We have established a preclinical model in which biocompatible ceramic scaffolds are populated with primary human bone marrow stromal cells (BMSCs) and can be implanted in mice, to engineer a mineralized bone-like matrix. This system simulates the histology, cellular composition and functional properties of the human bone marrow (BM)/bone, including the tumor-stromal interactions operating in hematologic malignancies, such as multiple myeloma (MM). Indeed, this humanized BM-like model has allowed us to engraft, expand and even serially transplant in vivo, patient-derived xenografts (PDX) from standard- and high-risk cases of MM, several types of acute and chronic leukemias, and myelodysplasia. PDX samples from these neoplasias typically have limited engraftment in conventional murine models. Our recent studies also show that PDX cells engrafted in this humanized model exhibit no substantial skewing of their immunophenotype, self-renewal properties or clonal architecture; while the outcomes of their preclinical treatments with conventional or investigational therapies are concordant with clinical responses to these treatments for the respective patients. We will extend this experience and systematically quantify the degree to which this humanized BM- like model allows human MM PDX cells to recapitulate their clinical features in this preclinical setting. To address this objective, we wil specifically examine whether MM PDX samples introduced in this humanized BM-like model exhibit higher engraftment rates, improved concordance of their preclinical responses to the respective patients' clinical treatment(s); and preservation of their clonal architecture, compared to when they are implanted in murine tissues (Specific Aim 1). We will also examine whether these quantitative metrics of translational relevance are improved when MM PDX samples are introduced in our model within scaffolds humanized with patient-derived BMSCs, compared to scaffolds with healthy donor-derived BMSCs (Specific Aim 2). Our studies will provide the framework for a new paradigm in PDX models, whereby biocompatible scaffolds functionalized with human stromal cell populations can help establish genotypically diverse, easily expandable patient-derived pre-clinical models to study the biology of human tumors and their therapeutic response in an individualized, patient-specific, manner.